Erich Elsen
18 papers · 2016–2022 · 4 conferences · across top CS/AI conferences
Achievements
Jump to papers ↓+10 more ↓ Show less ↑
π Academic Marathon (6) π Conference Polyglot (4) π Interdisciplinary Bridge π§ Keyword Pioneer π Cross-Pollinator (7)
π
Cross-Pollinator
(7)
π
Renaissance Researcher
(6)
πΊοΈ
Taxonomy Completionist
(23)
π₯
Mega-Team
(69)
π
Triple Crown
π₯
Unstoppable
(5)
β‘
Prolific Year
(5)
π
Century Club
(18)
ποΈ
Keyword Collector
(53)
π
Conference Pioneer
Conferences
ICML (9)
ICLR (6)
NIPS (2)
CVPR (1)
Top co-authors
Keywords
parameter efficiency
(3)
model compression
(3)
waveform generation
(2)
scaling law
(2)
sparse neural network
(2)
sparse training
(2)
neural network pruning
(2)
speech synthesis
(1)
stochastic gradient descent
(1)
deep reinforcement learning
(1)
computational efficiency
(1)
batch processing
(1)
recurrent neural network
(1)
deep neural network
(1)
efficient inference
(1)
feed-forward network
(1)
document retrieval
(1)
mixture of expert
(1)
end-to-end learning
(1)
question answering
(1)
Papers
The State of Sparse Training in Deep Reinforcement Learning
ICML 2022
Step-unrolled Denoising Autoencoders for Text Generation
ICLR 2022
Improving Language Models by Retrieving from Trillions of Tokens
ICML 2022
Unified Scaling Laws for Routed Language Models
ICML 2022
An empirical analysis of compute-optimal large language model training
NIPS 2022
Practical Real Time Recurrent Learning with a Sparse Approximation
ICLR 2021
End-to-end Adversarial Text-to-Speech
ICLR 2021
Top-KAST: Top-K Always Sparse Training
NIPS 2020
Fast Sparse ConvNets
CVPR 2020
High Fidelity Speech Synthesis with Adversarial Networks
ICLR 2020
Rigging the Lottery: Making All Tickets Winners
ICML 2020
On the Generalization Benefit of Noise in Stochastic Gradient Descent
ICML 2020
Enabling Factorized Piano Music Modeling and Generation with the MAESTRO Dataset
ICLR 2019
Parallel WaveNet: Fast High-Fidelity Speech Synthesis
ICML 2018
Mixed Precision Training
ICLR 2018
Efficient Neural Audio Synthesis
ICML 2018
Deep Speech 2 : End-to-End Speech Recognition in English and Mandarin
ICML 2016
Persistent RNNs: Stashing Recurrent Weights On-Chip
ICML 2016